Founders hire a LinkedIn ghostwriter for the posts that ship. Twelve a month, in your voice, on your lane. That's the invoice line, and it's a fair description of the deliverable.
It is not a complete description of the service. Because a meaningful share of the value we deliver shows up as posts that never got written — the idea a founder brought to a voice sync with real momentum behind it, that we pushed back on, and that quietly didn't happen.
Nobody puts that on a scope of work. It's hard to price and impossible to screenshot. But if you're evaluating ghostwriters, the willingness to say no is one of the few things that reliably separates an editor from a typist. A writer who executes everything you ask is cheaper to work with in every week except the ones that matter.
Here are the five we decline most often.
1. The score-settling post
A supplier burned you. A former agency ran your account into the ground. A competitor is running ads that misrepresent their product. You want to write about it, and you're careful — no names, just the pattern.
We understand the impulse and we usually kill the post.
Not for politeness. Because unnamed grievance content reads as unstable to the exact person you want reading it. A stranger scrolling past can't verify who was in the wrong. What they can see is a founder who takes business conflict public. If that stranger is an acquirer, a prospective 3PL, a retail buyer, or someone deciding whether to bring you a deal, you have just given them a data point that costs you more than the post earns.
There's a real version of this that we will write: the mechanism, with the emotion removed and the specific party unidentifiable. "Here's the contract clause that lets a 3PL raise your rates mid-term, here's what to ask for instead" is useful, forwardable, and makes you look like someone who has been around. "Some partners just don't operate in good faith" is a subtweet with a business logo on it.
The test we apply: would this post still be worth writing if the person it's about never saw it? If the answer is no, the post isn't content. It's a message, and it should be sent as one.
2. The borrowed take with no receipt
A founder reads a thread, likes the argument, and wants a post built on it. Usually there's a number attached. Return rates are up 40%. Rufus drives some enormous share of discovery. A platform change is coming on a specific date.
Our first question is where the number came from. Often the honest answer is "I saw it somewhere."
We won't publish that, and the reason is narrow and practical rather than moral: your audience contains the one person who can check. That's the operator you actually want — the one running a real catalog, who reads your post, recognizes that the figure doesn't match what they see in their own account, and quietly reclassifies you as someone who repeats things. They don't comment. They don't correct you. They just stop treating your specifics as reliable, which is the entire asset you've been building.
What we do instead is one of three things: publish the mechanism without the magnitude (the part you actually know), attribute the uncertainty out loud, or replace the borrowed number with something from your own account. The third is almost always the strongest post, and it's the one the founder didn't think of because the borrowed stat felt like the interesting part.
3. The announcement nobody asked for
We're hiring. We hit a milestone. We were named to a list. Happy Monday, here's a thought.
These aren't harmful. They're just withdrawals from an account you spend the rest of the month funding. Every post occupies a slot, and slots are the scarce resource — most founders can sustainably ship three a week, which is roughly 150 a year. Spending eight of them on internal news is spending 5% of your annual distribution on content that helps a stranger with nothing.
The exception is real, and we'll take it: an announcement carrying an operational lesson. "We're hiring our first ops person, here's the sequence of failures that made it obvious we needed one, and here's the number that finally moved me" is a post. "We're hiring, DM me" is a job board listing wearing a founder's face.
4. The off-lane hot take
Six months into an engagement, a founder gets bored of their own subject and wants to write about hiring philosophy, or the state of the economy, or an AI take that has nothing to do with their catalog.
Sometimes the answer is yes. Usually the answer is a narrower version of what they actually want.
The cost of going off-lane isn't reputational, it's structural. The platform has spent months learning who to show your content to, and that learning is topic-shaped. A broad post doesn't just underperform — it distributes to a different, less qualified room, and the signal you send about what your account is for gets muddier. Meanwhile broad takes are agreeable, so they collect easy engagement, and the vanity numbers stay fine while the inbound goes quiet. By the time anyone connects those two facts, it's been a quarter.
So we redirect. Boredom with a lane is almost never a request to leave it — it's a signal to go a level deeper into it, or to change format, or to write for a different reader inside the same subject. Same knowledge, different post. The founder gets the variety they were actually asking for and keeps the ground they've earned.
5. The post that spends a number it shouldn't
Ecommerce founders have unusually good raw material and unusually exposed raw material. Margins, landed cost, return rates, supplier terms, unit volumes — the specifics that make the content work are the same specifics your competitors, your suppliers, and your retail partners would like to have.
We won't publish a number that pairs a live SKU with information a competitor could act on, and we won't publish unit volumes attached to a category and a price point, because that reconstructs your reorder size and your supplier leverage in about ninety seconds.
But the useful part of this one is that the answer is almost never "don't write the post." It's "write it one rung down." Ratios without volumes. Bands instead of exact figures. Direction and magnitude instead of the raw number. Mechanism with no number at all.
And here's the thing founders consistently get backwards: the version without the number is often the stronger post. A number proves you were there. A mechanism proves you understood it. Anyone can quote a figure; only someone who lived it can explain why it moved.
Why "yes to everything" is the expensive option
A ghostwriter who never pushes back is easy to work with. The engagement feels frictionless. Every idea you bring becomes a post, and the calendar stays full.
That arrangement is also the one where you're paying a professional rate for transcription. The judgment is the product. The writing is the part that's become commoditized — a competent draft is available to anyone with a subscription and twenty minutes. What isn't available is someone who has watched a few hundred posts land across a portfolio of operators and can tell you, before you publish, that this particular one is going to cost you.
We're not describing a veto. Founders own their profiles and their opinions, and we've published plenty of posts we argued against and were wrong about. What we're describing is the obligation to say so first, with a reason, once — and then execute properly if the founder still wants it.
If you're vetting a ghostwriter, ask them directly: what have you talked a client out of publishing, and what was your reasoning? A writer with a real answer will have one immediately, and it'll be specific. A writer who says they'd never overrule a client's judgment is telling you something accurate about what you'd be buying.
FAQ
Isn't this just the ghostwriter imposing their opinions on my brand? No — it's the ghostwriter surfacing a cost before you pay it. You decide. The difference between an editor and a gatekeeper is that an editor makes the case once and then does the work either way.
What if I disagree and want it published anyway? Then it gets published, properly. The failure mode isn't a founder overruling their writer. It's a writer who never raised the concern, or one who raises it and then delivers a deliberately weak draft to prove a point.
How often does this actually come up? For most clients, a handful of times a year — and it clusters in months one and two, and again around month six when lane fatigue shows up. It's not a constant negotiation. It's a few high-consequence moments where having someone outside your head is worth the whole retainer.
Doesn't declining posts slow down my output? It's usually neutral or positive, because the redirect produces a replacement. The post we kill is almost always swapped for a stronger one from the same conversation — the aside the founder mentioned on the way to the idea they were excited about.
If you're an ecommerce founder who wants a writing partner with editorial judgment rather than an order-taker, that's what we do. We work exclusively with ecommerce founders and Amazon operators, which is the only reason we're able to tell you when something's a bad idea.